Opponents to the moratorium on state AI laws that's pending before the Senate fail to realize that we've been here before. Under the Articles of Confederation, the nation found itself economically constrained by a governance structure ill-suited to new economic realities. The severity of the mismatch between yesterday's approach to government and tomorrow's economy was not lost on the Founders. Here's how James Madison explained the situation to Thomas Jefferson:
"The States are every day giving proofs that separate regulations are more likely to set them by the ears than to attain the common object. When Massachusetts set on foot a retaliation of the policy of Great Britain, Connecticut declared her ports free. New Jersey served New York in the same way. And Delaware, I am told, has lately followed the example in opposition to the commercial plans of Pennsylvania. A miscarriage of this attempt to unite the states in some effectual plan will have another effect of a serious nature….I almost despair of success."
Madison and our nation's founders acted on this concern by recalibrating the respective authority of states and the federal government to dictate interstate commerce. The Commerce Clause empowered Congress to lead in shaping a cohesive economic strategy, while preserving the authority of states to exercise their core police powers.
Today, we again find ourselves on the precipice of a new economic era—one that also exposes tensions between the proper distribution of power among states and the federal government. AI has the potential to usher in significant economic growth, albeit with notable short-term disruptions to current markets. Existing AI systems, which we will likely regard as primitive in just a few years, have demonstrated the capacity to accelerate scientific discoveries, streamline production processes, and uncover new products. Conservative estimates of AI-induced GDP growth land around 1.8% over the next ten years. PWC, on the other hand, forecasts 14.5% growth in the next five years across North America. A more holistic assessment reveals that AI’s promise carries far more than just financial considerations. Hospitals have turned to AI to augment the detection of diseases. Schools have leveraged AI to aid in literacy initiatives. And, of course, companies like Waymo have made autonomous vehicles a reality in many communities thanks to AI.
A slew of even well-intentioned state regulations introduces a despair similar to Madison’s—a fear that we will fail to realize a brighter future as a result of separate laws and norms thwarting these positive AI use cases.
The justification for a national approach to AI is even more necessary here than in the trade context that prompted Madison’s concerns. Whereas correcting Massachusetts’ trade policies was a relatively straightforward task, unraveling a state’s premature AI policy is a much harder task. If New York passes the RAISE Act, for example, it will result in the creation of a number of vested interests that will fight any deregulatory efforts. The auditors called for in that Act will create substantial regulatory friction out of self-preservation. Likewise, incumbent AI labs that benefit from higher barriers to entry created by state laws will have a similar incentive to push back against reforms.
What’s more, given the nascent stage of AI development, premature state AI regulations may cause significant and irreversible downstream effects on AI development and diffusion. Though many state laws are intended to apply to a narrow stage of large models, the overall chill in AI investment introduced by complex and varied state laws may hinder startups from getting into the space. This may delay or even deny the realization of some of the most promising use cases of AI.
Advocates of state regulation contend that their laws will have limited or no effects beyond their borders. That’s a highly questionable argument. The essential inputs to AI development — data, energy, and talent — need to freely flow across borders to encourage AI development. Just as one state’s trade policies impacted the national economy under the Articles of Confederation, one state’s interference in these critical inputs can alter the broader AI market.
Advocates for state-level solutions often point to the current absence of a robust national AI regulatory framework as justification for their actions. However, we should view this not as a void, but as an opportunity for crucial deliberation. The current lack of a comprehensive federal AI act is, in fact, a deliberate pause—a feature, not a bug. It affords us the necessary time for a truly national conversation, engaging diverse voices and perspectives, to understand AI's complexities and implications for our society.
This public discourse has not yet occurred--the “AI Insights Forums” hosted in 2023 by then-Senate Majority Leader Schumer were generally closed to the public and made up of a small cluster of tech executives and researchers with highly speculative assessments of the future of AI. Now’s the time for town halls, educational initiatives, and other gatherings. These participatory and informational sessions are a prerequisite for developing thoughtful, democratically informed federal policies that will have lasting positive impacts, rather than rushing into laws with potentially unforeseen consequences.
It’s also important to point out that the moratorium preserves the role for states intended by the founders. They can still enforce general consumer protection laws that will shield the public from AI’s very real potential harms.
The founders recognized that certain policy questions were best left to the federal government. Their conviction on that front led them to ditch outdated conceptions and frameworks. On the verge of a new technological era, we, too, must resist the temptation to let yesterday's assumptions shape decisions that reach beyond any state's borders and will have very real impacts on national well-being.
All views and opinions expressed in this article are those of the author and not necessarily those of The National Law Review.